Developing a New Classification of Urban and Rural Areas for
Policy Purposes – the Methodology
Peter Bibby,
School of Town and Regional Planning, University of Sheffield
and
John Shepherd,
School of Geography, Birkbeck College,
Director, Defra Rural Evidence Research Centre
1.
Introduction
1.1
In 2001 the Office of the Deputy Prime Minister (ODPM) – then the
Department of Transport Local Government and the Regions (DTLR) –
commissioned a wide-ranging review of the definitions of urban and
rural areas in use for policy purposes and statistical reporting. The
review covered both England and Wales and involved consultations
with over twenty-five Government Departments and sections within
them. 1
1.2
The need for such a study had been recognised for some time and was
reinforced during data gathering and analysis for the Urban and Rural
White Papers. 2 The review had five main objectives including the need
to identify those policies that required definitions of urban and rural
areas and to suggest a ‘core’ set of definitions that met a wide range of
policy needs. Importantly, the review was also to ‘… identify any new
techniques that could better meet established and anticipated needs.’ 3
1.3
The review recommended that the ‘core’ definitions of ‘urban’ and
‘rural’ should, in the medium term, comprise the DTLR (now ODPM),
1991 ‘urban area’ boundaries with their census-based populations 4 and
the Countryside Agency’s administrative area classification of urban
and rural local authority districts and wards. However, in the rural
domain, both the urban areas and the administrative area definitions
1
A Review of Urban and Rural Definitions, Project Report, 2001,
http://www.statistics.gov.uk/geography/downloads/Project%20Report_22%20AugONS.pdf.
2
Our Towns and Cities: the Future, Delivering an Urban Renaissance Cm 4911 TSO 2000,
Our Countryside: the Future. A Fair Deal for Rural England, Cm 4909, TSO 2000
3
op cit para 1.1.2
4
Now updated to 2001
1
had a number of drawbacks, especially in relation to evolving rural
policy on service delivery. The review recommended that a clearer,
more comprehensive approach to rural area definition was needed. In
essence this would involve the extension of the ‘land use’ approach
that underlay the urban areas definition in order to identify, define and
derive populations for the small towns, villages, hamlets and isolated
dwellings that made up the settlement pattern of rural areas.
1.4
In early 2002 five Government bodies, namely the Department for
Environment, Food and Rural Affairs (Defra), the Office of the Deputy
Prime Minister (ODPM), the Office for National Statistics, the Welsh
Assembly Government and the Countryside Agency, formed a
consortium to commission a new definition of urban and rural areas
using the approach recommended in the review report.
1.5
The group charged with implementing and validating the new definition
comprised the South East Regional Research Laboratory (SERRL), at
Birkbeck College, the School of Town and Regional Planning at
Sheffield University, the School of Computing at the University of
Glamorgan and Geowise Ltd of Edinburgh. This note describes the
methodologies used and the key decisions taken in developing a new
definition and classification of smaller urban areas and rural
settlements in England and Wales.
2.
Abstract: The New Definition of Rural Places
2.1
The remit for producing a new rural/urban definition stated that this
should apply to those places which lay outside Census Urban Areas
with a population of 10,000 or more. 5 The methodology for defining
rural places described here is thus applied to settlements otherwise
described as ‘Urban Areas’ with between 10,000 and roughly 1,500
population. However, the definition of ‘rurality’ with which we are
concerned reaches much further down the settlement hierarchy to
small villages, hamlets and isolated dwellings.
2.2
The identification of rural settlements is derived from a grid covering
England and Wales with some 35 million cells each of 1ha. Individual
residential addresses are captured where they occur within this grid
forming a pattern of household densities that is used as a proxy for
residential land use at a high degree of resolution. Residential densities
are then averaged for each cell using a set of varying radii around each
cell. The result is the creation of a ‘density profile’ typifying settlements
and enabling, via a set of rules, a classification of settlement types.
2.3
The next stage is to relate rural settlements to Census Output Areas
and wards and to classify them by settlement type. This classification is
based upon the proportion of the population within each Output
5
See Census 2001 Key Statistics for Urban Areas, Office for National Statistics 2001
2
Area/ward in settlements of various kinds. Using the same general
approach, residential densities are also averaged at a series of much
larger geographic scales to give a ‘context’ measure for settlements
reflecting the wider ‘sparsity’ of the population. The classifications of
Output Areas/wards by settlement type and context are then brought
together to create a two-level classification of rural areas as follows:
Rural
Sparse
Town and fringe
Village
Less sparse
Dispersed
Town and fringe
Village
Dispersed
3.
Some Urban-Rural Distinctions
3.1
Although this is not the place for a discussion of whether, at the
cultural, social or economic levels, it is any longer valid to try to
distinguish between the ‘urban’ and the ‘rural’, 6 some comments in this
area may be helpful to understanding the approach to definition and
classification undertaken here.
3.2
There are at least three senses or dimensions that can be ascribed to
the word ‘urban’ and, by extension, to the word ‘rural’. One is
concerned with an apparently simple distinction between land that is
built over and land that is not built over. Usually, however, this
dimension is also associated with some threshold population size,
which might serve to distinguish between larger (urban) and smaller
(rural) settlements. 7
3.3
A second dimension to the term ‘urban’ is concerned with the wider
context of particular, physically defined, settlements. For example, we
might speak of an urban centre within an essentially ‘rural’ sub-region
or county, leading to the apparently contradictory notion of a ‘rural
town’. Context might, however, also relate to the broader settlement
structure in which a rural place is located, for example, set within a
landscape of nucleated villages or a mix of hamlets and isolated
dwellings.
3.4
The contextual sense of the term ‘urban’ is of potential significance to
policy because it might indicate the costs of delivering key services
such as health and education. The importance of context is, for
6
A discussion of matters related to this issue appears in the Review report chapters 5 and 6.
7
This is the approach taken, for example, in international definitions of urban areas. See
United Nations. World Urbanization Prospects, The 1999 Revision, United Nations publication
E.01.XIII.11, Chapter. VIII.
3
example, reflected in the inclusion of measures of population ‘sparsity’
in the local government revenue support grant formula.
3.5
Thirdly, the term ‘urban’ has long been used to denote economic
separation from the land as a direct source of income or wealth
generation. This, the functional dimension of ‘urban’ is at least equally
important as the other two and, indeed, is dominant within the social
sciences. Moreover, different types of settlement are associated with
different services, although this relationship can change over time as
the overall organisation of economy and society changes.
3.6
The new definition and classification of urban and rural areas reported
on here places its main emphasis on the morphology of rural
settlements (i.e. their physical form) and the wider geographic context
of such settlements. This approach ensures that the focus remains
clearly on the most enduring – physical - aspects of settlement.
4.
The Basics of the Classification.
4.1
As noted above, the definition and classifications presented here build
on the recommendations of a review of urban and rural definitions as
subsequently accepted by Government. The report concluded that it
was appropriate for most policy purposes to employ the ‘physical
settlements’ definition as represented by the ODPM defined ‘urban
areas’ and to treat those with more than 10,000 people as ‘urban’. 8
Given this, all other settlements are treated as part of a ‘rural’ domain.
4.2
The process of definition described here has as its initial ‘raw material’
all settlements of whatever size below 10,000. 9 Identifying and locating
these smaller settlements requires information at a very high level of
geographic resolution. This is provided by Royal Mail’s ‘Postcode
Address File’ (PAF) as packaged in its Address Manager ® product.
Amongst other things, PAF contains the postal addresses of premises
together with a 10m resolution OS grid reference for the unit postcode
allocated to each address.
5.
Clustering Postal Addresses
5.1
The starting point in the process of creating a representation of rural
settlement – what we have called the ‘underlying settlement
classification’ – is the grouping of every postal address on the basis of
the 1 hectare (100m x 100m) cell within which it falls.
8
Note that this is irrespective of the contextual or functional characteristics of urban areas
with more than 10,000 population i.e. what are regarded in the Rural White Paper as larger
‘market towns’ (op. cit. Chapter 7).
9
i.e. even though the Urban Areas definition itself relates to places with between 10,000 and
c 1000 population.
4
5.2
This process is illustrated in Figure 1, which shows, by means of the
pecked lines, some urban areas represented in the ‘core’ definition of
‘urban’. The larger central urban area (Canterbury) has more than
10,000 population and would thus not be included in the definition
analyses described here. The remaining settlements are below this
threshold. The variously coloured 1ha cells indicate the number of
postal addresses allocated to each cell. Note the clustered and isolated
coloured 1ha cells outside the pecked lines. These, along with the
small urban areas, are the focus for an analytical process designed, in
the first instance, to classify rural settlement on the basis of settlement
type or ‘morphology’.
Figure 1: Household Densities in 1ha Cells
5
6.
Constructing the Underlying Settlement Classification.
6.1
As noted above, the allocation of residential addresses to a regular grid
immediately allows examination of the density of households at the 1ha
cell level. Using the 2001 Second Quarter version of PAF means the
grid is virtually coterminous with information on the distribution of
households from the 2001 Census. The relationship between the two
for part of England is shown Figures 2 (a) and (b).
Figure 2a: Distribution of Households from the 2001 Census
of Population
6.2
Although the 1ha grid makes the perception and understanding of (in
this case) address densities more transparent, such densities are, in
fact, partly a function of the scale at which they are measured. As
areas are extended, for example, more areas of open space may be
included and average densities will decline. Importantly, we can make
use of this property - which gives different typical densities at different
scales - in order to identify and classify rural settlements. Here we
make use of the term ‘density profile’ which is used to refer to a series
of density measures focused on a given 1ha cell but calculated at
different scales. Furthermore, different morphologies or settlement
forms can be shown to have different typical density ‘profiles’.
6
Figure 2b: Distribution of Households from PAF 2001 Second
Quarter
6.3
Consider, for example, a situation where about 50 houses stand on a
relatively small piece of land, perhaps constituting a small village. If
one were to calculate density over a broader area centred on the 1ha
cell containing the village (say, for 200m radius around the centre of
that cell), and there were no other houses outside that cell, then the
density measure calculated over this wider area would be 4 dwellings
to the hectare. If one were to measure density over an area 400m
around the cell, the density would fall by a factor of four, to 1 dwelling
per hectare. 10
6.4
The rate at which density changes away from the ‘focus’ cell is a
function of local settlement structure. Thus in a conurbation, where
densities are sustained at (say) 30 dwellings to the hectare over a
broader area, such falls will not occur, whereas for a village in an area
of hamlets and isolated dwellings, the density ‘fall-off’ will be marked.
6.5
‘Density profiles’ can thus be created using a series of different area or
‘window’ sizes. In other words, density profiles can be created by
calculating densities at a series of fixed scales - in our case 200m,
400m, 800m and 1600m - around each cell (Figure 3). Actual
10
Clearly, at the other extreme i.e.in a city, this would not be the case, as higher densities
would be maintained over larger areas.
7
settlement patterns, of course, are not composed solely of compact
villages surrounded by un-populated agricultural land. In practice, the
decline of density with distance is less acute. Typically, a ‘village’ as
defined here would have the following properties: a density of greater
than 0.18 residences per hectare at the 800m scale, a density at least
double that at the 400m scale and a density at the 200m scale at least
1.5 times the density at the 400m scale.
Figure 3: Areas for Calculating the Grid Density
Gradient
Dark Green : area for
calculating the 200m density
Pale Green : area for
calculating the 400m density
Pink : area for calculating the
800m density
Beige : area for calculating
the 1600m density
6.6
To illustrate this, consider the example of Great Rissington in
Gloucestershire, a village of about 360 dwellings. At the 800m scale
the typical density for a hectare cell in the village is 0.73 dwellings to
the hectare; at the 400m scale the corresponding density is 2.94 and at
the 200m scale it is 11.08. (See Figure 4). Larger settlements defined
here as ‘towns’ also have a distinct profile. This can be illustrated by
reference to Henley-in-Arden in Warwickshire. The greater physical
and population size of this town is reflected in its higher density at the
1600m scale which is the critical scale for the identification of ‘rural’
towns. At the 400m scale Henley, typically, has a density of 18
dwellings to the hectare, falling to 7.3 at the 800m scale and to 2.0 at
the 1600m scale (Figure 5).
8
Figure 4 A Compact Village Identified by Density Profiles
Stow
Typical household densities
(dwellings/ha)
Bourton
360 dwellings
800m scale : 0.73
400m scale: 2.94
200m scale: 11.08.
Northleach
Great Rissington
Burford
See Figure 6 for colour codes to settlement type.
9
Figure 5 A Rural Town Identified By Density Profiles
Redditch
Henley
in Arden
Typical household densities
(dwellings/ha)
c 1400 dwellings
Alcester
1600m scale : 2.1
800m scale: 7.3
400m scale: 18.3
Stratford
See Figure 6 for colour to codes settlement type.
10
6.7
The use of ‘density profiles’ at the sorts of geographic resolution
discussed here can identify a whole range of elements within the urban
and rural settlement structure. Such elements include not only a clear
delineation of the boundaries of large and small urban areas where
these follow residential land uses, but also such features as the ‘urban
fringe’ (where there are abrupt changes of density between scales),
nucleated villages and their ‘envelopes’ and areas of scattered
dwellings. Areas of higher density dispersed settlement around cities
and towns also have a distinct ‘peri-urban’ density profile.
6.8
In fact, the underlying settlement classification allows for the
identification of nine morphological types including, for example, ‘urban
fringe’, ‘town’, ‘village’ and ‘hamlet’. Seven of these features are shown
in Figure 6 – those not present in this area are hamlets and isolated
farms. The numerical outcome of applying the procedure to the 1ha
cells for England and Wales is shown as a set of overall average
densities for each identifiable rural settlement feature in Table 1. Annex
1provides then rules for identifying settlement morphology.
Table 1: Measured Density Profiles for Settlement Forms
Settlement Form
Small town
Urban fringe
Village
Peri-urban
Village envelope
Village envelope (in peri-urban)
Hamlet
Scattered dwellings
Urban Areas (above 10k)
Density of Residential Delivery Points (mean)
At 200m
At 400m
At 800m
At 1600m
8.23
8.99
8.29
5.59
6.46
7.21
5.90
4.68
3.81
2.28
0.83
0.58
0.30
0.59
1.57
2.80
0.94
1.15
1.31
0.59
2.96
3.27
1.81
2.13
0.65
0.21
0.13
0.20
0.39
0.17
0.15
0.23
16.09
15.17
13.78
11.89
Note: it is important to recognise that these data are the outcome of applying the
density profile measurement procedure and hence take account of the particular
environs of settlements.
6.9
11
Finally, use of the textual descriptions of addresses within the PAF
allows both historic and functional elements of settlements to be
identified. Thus within the broad category of ‘dispersed settlements’,
hamlets were identified in the developmental stages of the definition.
By applying natural language processing, PAF is used to identify
farmsteads and then, in the tradition of rural settlement analysis
exemplified by Roberts (1996), 11 to identify hamlets and larger
settlement units. Clusters of 3-8 historic farmsteads within 250m of
each other were identified as ‘hamlets’.
B K Roberts, Landscapes of Settlement, London, Routledge, 1996
11
Figure 6: A Range of Rural Settlement Features Identified by Density Profiles
Ascott
Moreton in Marsh
Hook Norton
Long Compton
Urban Area
Peri urban
Longborough
Evenlode
small town
urban fringe
Chipping Norton
Stow on the Wold
village
Churchill
Hyde Hill
village envelope
Dean
scattered dwellings
Spelsbury
Charlbury
12
7.
Adding Context
7.1
‘Context’, as noted above, refers to the broader setting in which
settlements (including isolated dwellings) are located. In this sense
context may be interpreted in general terms as the wider accessibility
of a settlement, the sparsity of population within a broad area and, in a
general way, the potential costs of overcoming distance to supply that
settlement with various public and private services.
7.2
In just the same way as they have been used to identify settlement
types, density profiles can be used at much larger scales to
characterize aspects of accessibility and population sparsity. In our
case this involves calculating for every 1ha cell the density of
households across areas of 10,000m, 20,000m and 30,000m centred
on that cell. 12 Each of these context measures for a particular 1ha cell
might be thought of as deriving from a household density map at the
stated scale. Maps for the 10,000m and 30,000m scales are shown in
Figure 7.
7.3
On the basis of these measures it is possible to identify areas where
population is ‘sparse’ at the particular scale. By assigning these
measures to 2001 Census Output Areas and focusing on the sparsest
5 percent in each case, three indicators of ‘sparsity’ are obtained. The
map of the grids that meet this criterion at all three scales is shown in
Figure 8.
7.4
For definitional purposes (see the section headed ‘Output Areas’
below), the focus has been on areas whose population might be
considered ‘sparse’ at all three scales. However, a clear distinction
might be made between areas ‘sparse’ at all three scales (such as
Central Wales and parts of North Devon), and those ‘sparse’ only at the
10km scale such as the Cotswolds and the White Peak of Derbyshire).
This distinction has potential for further application and, as a result of
comments made in the validation procedure, is being explored.
12
These scales were arrived at so as to broadly typify (at the 10,000m range) general
commuting distances and, at the larger scales, the range of supply of ‘high level’ rural
services such as ambulances or fire fighting vehicles. The distances selected were the result
of discussion with the Project Board and empirical experimentation.
13
Figure 7: Household Densities Calculated at 10,000m and 30,000m.
(b) 30,000m
(a) 10,000m
14
Figure 8 The Combined Context (Sparsity) Map
Dark Bue: ‘sparse’ at all
three scales
Mid Blue: ‘sparse’ at
30,000m and 20,000m
Light Blue: ‘sparse’ at
10,000m
15
Considering Function
8.1
As part of the work of producing a new urban/rural definition, function
was also considered. ‘Function’ here refers to the economic character
of settlements or, again and more precisely, to the 1ha cells that
constitute settlements.
8.2
For this purpose, indicators were developed which distinguished cells
in accordance with the mix of residential and non-residential
addresses. This allowed the definition, for example, of dormitory
settlements. Moreover, using natural language processing, 13 an
assessment was made of the Standard Industrial Classification and
town and country planning Land Use Class of each postal address on
the basis of the premise described and the occupier name.
8.3
In this way, every 1ha cell was considered either to be characterized by
‘no businesses’, ‘farm businesses’, ‘tourist business’ or ‘other
businesses’. These indicators were then considered alongside the
morphological and context measures for possible inclusion as
indicators for use in producing urban and rural definitions.
9.
The Classification of Statistical and Administrative Units
9.1
Having classified individual cells, the next step is to categorise the
settlement characteristics of statistical and administrative units. This is
designed to complement the use of a range of statistical data,
particularly those from the decennial population census. The smallest
of the units to be classified are 2001 Census Output Areas. Output
Areas were introduced for the first time as a basis for publication of
results of the 2001 Census. 14 They are intended to represent compact,
socially homogenous areas, and are designed to nest within ward and
parish boundaries and hence within higher level administrative
geographies. 15
9.2
At a broader scale, wards are also classified. In each case, the
calculation of indicators for each statistical unit involves allocating the
1ha cells within various classes of settlement to the areas concerned.
Classification then depends on identifying the proportion of population
in each settlement type within the area concerned. Once all the
relevant indicators have been calculated, classification can then
proceed.
13
P Bibby, Maps from Words, International Journal of Geographic Information Science,
forthcoming
14
http://www.statistics.gov.uk/census2001/cn_40.asp
15
http://www.geog.soton.ac.uk/research/oa2001/
16
(a) Census Output Areas
9.3
At Census Output Area level, units are grouped into four morphological
types on the basis of their predominant settlement component:
•
•
•
•
urban
town and fringe
village, and
dispersed.
9.4
Output Areas are treated as ‘urban’ or ‘rural’ simply on the basis of their
geographic relationship to settlements of 10,000 or more population.
More specifically, where the majority of the population of an Output
Area lives within settlements with a population of more than 10,000
people, that Output Area is treated as urban. All other Output Areas are
treated as rural. This is superimposed on the underlying settlement
classification to form what is called here the ‘combined settlement
classification’.
9.5
Output Areas are given a sparsity score at 10km, 20km and 30km by
producing a weighted total of 1ha squares within an Output Area,
where the weights are the number of residential delivery points (rdp).
Output Areas are classified as ‘sparse’ if they fall within the sparsest 5
percent of Output Areas at all three scales and are classified as ‘less
sparse’ if they do not fall within this threshold. The fifth percentile ‘cutoffs’ for the application of this rule are shown in Table 2.
Table 2 Fifth Percentile Measures for Defining Sparsity
Category
Sparse at the 10km scale
Sparse at the 20km scale
Sparse at the 30km scale
Fifth Percentile
< 0 .3932 rdp per ha
< 0.41 rdp per ha
< 0.4224 rdp per ha
9.6
It should also be noted that Output Areas are classified by ‘hierarchical
privileging’, that is, if an Output Area has 50 percent by area of a
particular settlement morphology, this classification is used. However,
in the very small number of cases where an Output Area did not
contain a dominant morphological type then the largest settlement
character is ‘privileged’ with the Output Area classification.
9.7
Finally, classifications on the ‘morphology’ and ‘context’ dimensions are
combined as indicated in the ‘tree diagram’ (Figure 9) and mapped in
Figure 10.
17
Figure 9: The Proposed New Rural Classification (Output Area level)
RURAL
Less
Sparse
Sparse
Town and
fringe
Village
Hamlets
and
dispersed
Town and
fringe
18
Village
Hamlets
and
dispersed
households
Figure 10: Output Areas Classified
19
(b) Census Wards
9.8
Similar procedures are applied to wards. The design of electoral wards,
however, is such that very few are characterized by predominantly
dispersed settlement (in fact, only 0.5 percent). For this reason, only
three morphological categories are distinguished:
•
•
•
9.9
urban,
town and fringe, and
village and dispersed
Because context measures vary smoothly from 1ha cell to another,
there is little difficulty in estimating measures for wards that are
consistent with those measured for Output Areas, as illustrated in
Figure 11.
Figure 11: 2001 Census Wards Classified
20
10
Local Authority Districts
10.1
Broadly speaking the same classificatory principles can be applied at
larger geographic scales. Morphological classification of local authority
districts (LADs) is, however, much less straightforward. For this reason
we are not recommending, at this stage, a definitive binary (i.e.
rural/urban) classification of LADs, but simply note some issues for
further analysis.
10.2
The design of territories for local authorities tends to include a mix of
urban and rural areas (typically with a population of 100,000 or more).
Just as the dispersed settlement category disappears when moving
from the Output Area to the ward scale, a shift to the local authority
district scale involves the ‘collapse’ of most of the rural morphological
categories.
10.3
On the basis discussed above, three quarters of such districts appear
as ‘urban’ in morphological terms, whilst the remainder have no
predominant settlement form. This perhaps suggests that a much
simpler morphological classification is required at the LAD level. While
it might seem possible to distinguish those LADs that are
predominantly ‘urban’ in morphological terms from others, this might
appear problematic because a substantial number of LADs are
characterized by settlement patterns where the urban component
accommodates between 40 and 60 percent of the total population
(Figure 12).
Figure 12 Number of Local Authority Districts, Percentage of
Households in Urban Areas
120
102
100
80
Frequency
60
40
18
20
7
6
0
12 10
11 11
15
18
20
14
25 23
24 23
16 15
5
0
00
-1
95
5
-9
90
0
-9
85
5
-8
80
0
-8
75
5
-7
70
0
-7
65
5
-6
60
0
-6
55
5
-5
50
0
-5
45
5
-4
40
0
-4
35
5
-3
30
0
-3
25
5
-2
20
0
-2
15 5
-1
10
10
5-
5
0-
% Urban
21
10.4
Once again, however, the gradual nature of variation in the context
measures facilitates their application to the local authority district scale.
Only 14 local authority districts (3.7 percent), are ‘sparse’ at all three of
the measured scales. The resulting geography of districts in ‘sparse’
areas is shown in Figure 13. 16
Figure 13: Local Authority Districts with Sparse Population
on the Basis of all Three Density Scales
16
Recognising the potential significance of and need for a district level definition of ‘rurality’
the Minister of State for Rural Affairs has asked the Rural Evidence Research Centre at
Birkbeck College to explore the idea and present proposals.
22
Annex 1
Rules for identifying rural settlement morphology.
The settlement morphology element of the rural definition is derived from the
classification of 1ha grid squares according to a set of rules.
These rules were derived empirically and were developed from the typical
characteristics of residential building densities and from comparison of the
outcomes of the application of the rules with OS map sources.
The hectare square classification rules are applied to all cells in the order
given below. Hence a subsequent application of a rule identifying ‘fringe’ may
have different parameters but will identify the same settlement type.
Densities are calculated over a series of radii. The density over a 1600m
radius around a cell is referred to as the D1600 measure for that cell. The
terms ‘D800’, ‘D400’, and ‘D200’ are defined in a similar manner.
Rule 1:
If
D800 > 8
a grid square is deemed to form part of a rural town or urban area
Rule 2:
If
D800 > 4
and
D400 > 4
and
D800 < 8
and
D800 > 2.5*D1600,
a grid square is deemed to form part of a fringe.
Rule 3:
If
D800 < 8
and
D400 > 8
D800 > 2.5*D1600,
a grid square is deemed to form part of a rural town.
Rule 4:
If
D800 > 4
D400 > 4
and
and
23
D800 < 8,
a grid square is deemed to form part of a fringe.
Rule 5:
If
D800 < 8 and
D400 > 8,
a grid square is deemed to form part of a rural town.
Rule 6:
If
D800 > 0.18 and
D400 > 2*D800 and
D200 > 1.5*D800,
a grid square is deemed to form part of a village.
Rule 7:
If
D1600 > 1.0 and
D 400 > 1.5*D800,
a grid square is deemed to form part of a village envelope (in peri-urban).
Rule 8:
If
D1600 > 1.0,
a grid square is deemed to form part of a peri-urban zone.
Rule 9:
If
D1600 =< 1.0 and
D800 > 0.5,
a grid square is deemed to form part of a village envelope.
24